Model Details

Model Card for Olmo 3 7B RL-Zero IF

We introduce Olmo 3, a new family of 7B and 32B models both Instruct and Think variants. Long chain-of-thought thinking improves reasoning tasks like math and coding.

Olmo is a series of Open language models designed to enable the science of language models. These models are pre-trained on the Dolma 3 dataset and post-trained on the Dolci datasets. We are releasing all code, checkpoints, logs (coming soon), and associated training details.

The RL-Zero family of models is an experimental set of model for the scientific exploration of RLVR training.

For the other Olmo 3 RL-Zero models see:

Domain	Model	RLVR Dataset
Base Model	Olmo-3-7B
Math	Olmo-3-7B-RLZero-Math	Dolci-RLZero-Math-7B
Code	Olmo-3-7B-RLZero-Code	Dolci-RLZero-Code-7B
IF	Olmo-3-7B-RLZero-IF	Dolci-RLZero-IF-7B
General	Olmo-3-7B-RLZero-General	Dolci-RLZero-General-7B
Mix	Olmo-3-7B-RLZero-Mix	Dolci-RLZero-Mix-7B

For the core Olmo 3 models see:

Stage	[Olmo 3 7B Think]	[Olmo 3 32B Think]	[Olmo 3 7B Instruct]	[Olmo 3 32B Instruct]
Base Model	Olmo-3-7B	Olmo-3-32B
SFT	Olmo-3-7B-Think-SFT	Olmo-3-32B-Think-SFT	Olmo-3-7B-Instruct-SFT	Olmo-3-32B-Instruct-SFT
DPO	Olmo-3-7B-Think-DPO	Olmo-3-32B-Think-DPO	Olmo-3-7B-Instruct-DPO	Olmo-3-32B-Instruct-DPO
Final Models (RLVR)	Olmo-3-7B-Think	Olmo-3-32B-Think	Olmo-3-7B-Instruct	Olmo-3-32B-Instruct

Installation

Olmo 3 is supported in transformers 4.57.0 or higher:

pip install transformers>=4.57.0

Inference

You can use OLMo with the standard HuggingFace transformers library:

from transformers import AutoModelForCausalLM, AutoTokenizer
olmo = AutoModelForCausalLM.from_pretrained("allenai/Olmo-3-7B-RL-Zero-IF")
tokenizer = AutoTokenizer.from_pretrained("allenai/Olmo-3-7B-RL-Zero-IF")
message = ["Language modeling is "]
inputs = tokenizer(message, return_tensors='pt', return_token_type_ids=False)
# optional verifying cuda
# inputs = {k: v.to('cuda') for k,v in inputs.items()}
# olmo = olmo.to('cuda')
response = olmo.generate(**inputs, max_new_tokens=100, do_sample=True, top_k=50, top_p=0.95)
print(tokenizer.batch_decode(response, skip_special_tokens=True)[0])
>> 'Language modeling is  a key component of any text-based application, but its effectiveness...'

For faster performance, you can quantize the model using the following method:

AutoModelForCausalLM.from_pretrained("allenai/Olmo-3-7B-RL-Zero-IF", 
    torch_dtype=torch.float16, 
    load_in_8bit=True)  # Requires bitsandbytes

The quantized model is more sensitive to data types and CUDA operations. To avoid potential issues, it's recommended to pass the inputs directly to CUDA using:

inputs.input_ids.to('cuda')

Fine-tuning

Model fine-tuning can be done from the final checkpoint (the main revision of this model) or the base model.

We recommend fine-tuning with the open-instruct repository:

bash ./scripts/train/olmo3/rlvr_script.sh

You can override most configuration options from the command-line. For example, to override the learning rate you could launch the script like this:

bash ./scripts/train/olmo3/rlvr_script.sh --learning_rate=1e-3

For more documentation, see the GitHub readme.

Model Description

Developed by: Allen Institute for AI (Ai2)
Model type: a Transformer style autoregressive language model.
Language(s) (NLP): English
License: This model is licensed under Apache 2.0. It is intended for research and educational use in accordance with Ai2's Responsible Use Guidelines.
Contact: Technical inquiries: [email protected]. Press: [email protected]
Date cutoff: Dec. 2023.

Model Sources

Project Page: https://allenai.org/olmo
Repositories:
- Open-Instruct for DPO and RLVR: https://github.com/allenai/open-instruct
- OLMo-Core for pre-training and SFT: https://github.com/allenai/OLMo-core
- OLMo-Eval for evaluation: https://github.com/allenai/OLMo-Eval
Paper: [TBD]

Model Details

RLVR

reinforcement learning from verifiable rewards on the Dolci-RL-Zero-IF-7B dataset which consists of instruction-following queries.
Datasets: Dolci-RLZero-IF-7B

Bias, Risks, and Limitations

Like any base language model or fine-tuned model without safety filtering, these models can easily be prompted by users to generate harmful and sensitive content. Such content may also be produced unintentionally, especially in cases involving bias, so we recommend that users consider the risks when applying this technology. Additionally, many statements from OLMo or any LLM are often inaccurate, so facts should be verified.

Citation

@misc{olmo2025olmo3,
title={Olmo 3},
author={Team Olmo and Allyson Ettinger and Amanda Bertsch and Bailey Kuehl and David Graham and David Heineman and Dirk Groeneveld and Faeze Brahman and Finbarr Timbers and Hamish Ivison and Jacob Morrison and Jake Poznanski and Kyle Lo and Luca Soldaini and Matt Jordan and Mayee Chen and Michael Noukhovitch and Nathan Lambert and Pete Walsh and Pradeep Dasigi and Robert Berry and Saumya Malik and Saurabh Shah and Scott Geng and Shane Arora and Shashank Gupta and Taira Anderson and Teng Xiao and Tyler Murray and Tyler Romero and Victoria Graf and Akari Asai and Akshita Bhagia and Alexander Wettig and Alisa Liu and Aman Rangapur and Chloe Anastasiades and Costa Huang and Dustin Schwenk and Harsh Trivedi and Ian Magnusson and Jaron Lochner and Jiacheng Liu and Lester James V. Miranda and Maarten Sap and Malia Morgan and Michael Schmitz and Michal Guerquin and Michael Wilson and Regan Huff and Ronan Le Bras and Rui Xin and Rulin Shao and Sam Skjonsberg and Shannon Zejiang Shen and Shuyue Stella Li and Tucker Wilde and Valentina Pyatkin and Will Merrill and Yapei Chang and Yuling Gu and Zhiyuan Zeng and Ashish Sabharwal and Luke Zettlemoyer and Pang Wei Koh and Ali Farhadi and Noah A. Smith and Hannaneh Hajishirzi},
year={2025},
eprint={2512.13961},
archivePrefix={arXiv},
primaryClass={cs.CL},
url={https://arxiv.org/abs/2512.13961},
}